Tracking Anthropogenic Inputs Using Caffeine, Indicator Bacteria, and Nutrients in Rural Freshwater and Urban Marine Systems KELLY A. PEELER, † STEPHEN P. OPSAHL, ‡ AND JEFFREY P. CHANTON* ,† Department of Oceanography, Florida State University, Tallahassee, Florida 32306-4320, and Joseph W. Jones Ecological Research Center, Route 2, Box 2324, Newton, Georgia 39870 Our objective was to evaluate the hypothesis that measurements of caffeine, nutrients, and indicator bacteria can distinguish human versus non-human sources of surface water contamination in contrasting environments. A second objective was to determine if natural sources of caffeine were significant in unpopulated areas. Caffeine was measured in an isolated wetland, and a native plant source was identified. In two rural watersheds in southwest Georgia (U.S.), caffeine was detected in tributary creeks immediately below wastewater discharge sites and within towns. However, caffeine was not found in river main streams. Thus, although natural caffeine sources exist, background levels in stream drainage networks of these rural watersheds remained below detection. The presence of caffeine and elevated nitrate in streams was associated with anthropogenic inputs and population centers, whereas bacterial indicators did not correlate to these chemical indicators and appeared to have non-human sources. In contrast, caffeine in an urban coastal lagoon was generally linked to fecal coliform abundance. We observed sporadic relationships between caffeine and other water quality indicators, possibly due to differential rates of degradation. Creeks and bayous flowing into the lagoon contained the greatest caffeine concentrations and highest amounts of bacteria, nitrate, and radon, which is an indicator of groundwater discharge. Introduction Surface and groundwater contamination from anthropogenic sources has become widespread in the southeastern U.S. because of rapid population growth and a lack of adequate municipal infrastructure to keep pace with development. Microbial indicator organisms have been widely used to trace the origin of wastewater and determine the likelihood that the water poses a significant threat to human health. Fecal coliforms and enterococci are two of the most commonly used indicator bacteria (1, 2). Nutrient concentrations are often elevated in wastewater and can have ecological consequences. There are certain shortcomings with relying solely on indicator bacteria to assess risks associated with pathogenic microorganisms in water. The currently used groups of bacterial indicators are present in many animals, making distinctions between human and other sources difficult (2, 3, 4). Another problem is that fecal coliform abundance is not always consistent with the abundance of pathogenic viruses and human health risks (2, 3). In some instances, human pathogens are more persistent in environmental waters than are fecal coliforms. Furthermore, fecal coliforms can adapt and survive in sediments and can later be mobilized into the water column via disturbances (5, 6, 7). Methods for microbial source tracking in aquatic envi- ronments have been developed to distinguish animal from human sources. Such methods include monitoring for host specific organisms (viruses) or host related nucleic acid sequences (1-4, 8, 9). Considerable progress has been made in this arena, however, methods such as bacterial source tracking may be of limited geographic use. Expense, repro- ducibility, and standardization have also been problems for these approaches (4). Caffeine has recently been examined as a tool for assessing human impacts on aquatic systems (10-20). When con- sumed, caffeine is metabolized (21, 22), but a small amount (0.5-10%) of ingested caffeine remains intact when excreted (16, 18, 21). Most work in the past decade has focused on heavily polluted systems and efficiency of caffeine removal in sewage treatment plants (10, 13, 14, 19, 23-25). However, with improvements in technique (13, 16, 17, 26, 27) and lowered detection limits the scope of application has broadened to include stream, wetland, estuarine, and groundwater systems (11-20). In many instances, there appears to be an association between elevated caffeine concentrations and high population densities (10-12, 18). However, to definitively trace caffeine to humans, the potential for natural sources needs to be more carefully considered. While caffeine is present in more than 60 species of plants (28), few are native to the U.S. (16). One species (Ilex vomitoria or yaupon holly) found in the southeast U.S. is known to contain caffeine (29). The application of caffeine and other anthropogenic markers show promise for understanding wastewater con- tamination of groundwater in coastal zones, yet only a few studies have examined caffeine in groundwater (13). The added use of the geochemical tracer radon would be beneficial for understanding the distribution and transport of contaminants such as caffeine in groundwater. Radon is a dissolved gas whose presence in surface waters is frequently attributed to groundwater discharge (30). The objective of this work was to evaluate the hypothesis that measurements of caffeine, nutrients and indicator bacteria can distinguish human versus non-human sources of surface water contamination in contrasting environments. We further hypothesized that natural background concen- trations of caffeine would be low and examined this hypothesis in rural areas. In Sarasota Bay, an urbanized lagoon, we hypothesized that the presence of caffeine would correlate to bacterial and nutrient contamination. Materials and Methods Sampling Sites-Freshwater. Three sets of samples were collected during January and February 2004 from rural areas in south Georgia and north Florida (Figure 1). “Grab” samples were collected from the middle of the stream or wetland and all were obtained from below the surface to avoid the surface micro-layer. Separate water samples were also collected in * Corresponding author phone: 850-644-7493; fax: 850-644-2581; e-mail: jchanton@mailer.fsu.edu. † Florida State University. ‡ Joseph W. Jones Ecological Research Center. Environ. Sci. Technol. 2006, 40, 7616-7622 7616 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 24, 2006 10.1021/es061213c CCC: $33.50  2006 American Chemical Society Published on Web 11/09/2006